MRI video diagnosis and surgical therapy of soft tissue trauma to the craniocervical junction.

Abstract

We evaluated objective diagnostic methods for patients with possible upper cervical spine instability caused by trauma and correlated them with subsequent neurosurgical findings and outcomes. Between November 1995 and May 1998, we investigated 420 patients with functional magnetic resonance imaging (MRI) of the craniocervical junction. We evaluated the extracranial vertebral circulation by MRI angiography, with focus on the position of the dens and on the subarachnoid space during entire rotational maneuvers. We documented 72 cases (17.1%) of injuries to the alar ligaments that were accompanied by signs of instability. Twenty patients (4.8%) had a complete alar ligament rupture, and 52 (12.4%) had an incomplete rupture with coexisting instability. We referred these patients to a neurosurgeon. Surgery was eventually chosen for 42 patients (10.0%) with the intention of obtaining dorsal occipitocervical stabilization. The duration of time between the MRI evaluation and surgery ranged from 1 week to 1.5 years (me an: 3.5 mo). After the fifth postoperative day, almost all symptoms had disappeared. One year following surgery, 34 of the 42 patients (80.9%) still demonstrated successful fusion and an alleviation of their sensation of instability. Twenty-five of these patients (59.5 %)--all of whom were unemployed before surgery--were able to resume a professional activity. In the eight patients (19.0%) who still had a loss of stability during the second and 14th weeks, we noticed that there were some negative effects of rehabilitation. Six of these patients developed pseudarthrosis or osteolysis of their bone grafts during the first 3 months after fusion, and three required a repeat operation. We conclude that functional MRI with lateral tilting and rotatory evaluation is a useful tool for investigating craniocervical instability. For patients who are recalcitrant to following a program of conservative therapy, surgical stabilization of the craniocervical junction appears to be justified.

Introduction

Patients who experience an injury to their cervical spine after an acceleration trauma often present problems with respect to the correct diagnosis. Particularly difficult to recognize are injuries to the cervicocephalic area because there is currently a lack of objective diagnostic criteria. The craniocervical ligaments and fibrous capsules are not visible on plain radiographs. A widened or uneven atlantodental distance implies that the alar ligaments are disrupted or dysfunctional.

Approximately 25% of all patients with cervical distortion and injury to the soft tissues of the neck experience cervical and/or neck pain up to 4 to 7 years later--pain that requires continual orthopedic, manual, pharmacologic, or other therapy. [1,2] Several authors have postulated that cervical spine instability is a consequence of injury to the ligaments. [3-6] Conventional x-ray studies and functional computed tomography (CT) can be helpful in determining the various angles of anatomic markers in the spine. [4,7,8] However, the usefulness of these imaging studies is dependent on the degree of the patient's relaxation and do not always correspond with the degree of functional impairment. It is known that the atlanto-occipital plane and therefore the socle joint (C2 vertebral body) are especially vulnerable to indirect trauma. [9] Also, the horizontally oriented facet joints and capsules between the atlas and axis can be affected by accentuated axial rotation, which can injure the alar ligaments.

Since the introduction of functional magnetic resonance imaging (fMRI) on an open magnet, it has been possible to observe the functional condition of the ligaments and the atlantoaxial joints on a video loop. One study attempted to classify alar-atlantoaxial joint instability and the related regional injuries in order to better understand the result of biomechanical damage to the ligaments during overstretching. [10]

Saternus found that among 397 victims of high-speed trauma, 340 (85.6%) had evidence of upper cervical ligamentous lesions, while only 57 (14.4%) had bone fractures. [11]

Functional MRI video diagnosis does not focus on injuries to the ligamentous microstructure as does high-resolution MRI. Instead, it directly visualizes instability of the craniocervical junction. [12] It is generally accepted that increased axial rotation instability of the upper cervical spine can cause symptoms such as severe occipital headache and pain and tenderness in the adjoining neck muscles. Other concomitant complaints can include dizziness, tinnitus, paresthesia, visual disturbance, cognitive impairment, sleep disturbance, vegetative symptoms, inability to ride a bicycle, and darkness orientation disturbances. [13,14]

Patients and methods

Between November 1995 and May 1998, we studied 420 patients (228 females, 192 males), aged 17 to 55 years (mean: 37), who had a history of trauma involving the upper cervical spine. We performed fMRI with a 0.2 Tesla Magnetom Open imager (Siemens; Erlangen, Germany) that was equipped with a device that allowed for lateral tilting and transverse rotation of the cervical spine. The causes of these traumas included high-speed motor vehicle or automobile-pedestrian collisions (n = 371), falls from high elevations (n = 18), sports play (n = 16), and bicycle or motorbike accidents (n = 15). We excluded from our study patients younger than 17 years of age and patients who had open, penetrating spinal injuries, metabolic bone disease, ankylosing spondylitis, rheumatoid arthritis, or generalized connective tissue diseases.

The duration of time between the trauma and the MRI ranged from 4 months to 5 years (mean: 2.7 mo). Most patients had earlier undergone plain radiography, and some had undergone thin-section CT or high-resolution MRI under static conditions. Before undergoing fMRI, all patients underwent MRI angiography of the cervical vertebral arteries. When necessary, we monitored heart rate and respiratory function during the imaging process with an MR-fiberoptic pulse oximeter (Nonin 8600 FO, Mediquip; Kirchzarten, Germany). There was continuous, direct-view visual monitoring of the patient. Our decision to use an open magnet was based on our familiarity with clinical manual therapy--specifically, tilting the neck step by step to the right and left as a single investigation and rotating the neck to the right and left as a separate investigation. No anesthesia was necessary. Under our protocol, the procedure could be interrupted if the patient lost consciousness or developed a blockage in the upper cervical spine that co uld irritate the vertebrobasilar circulation. The details of the MRI parameters and characteristics have been reported elsewhere. [15] Circular surface coils of different diameters were used to improve the anatomic resolution at the target point. We obtained thin slices, mostly between 4 and 5 mm, that were oriented to the exact location of the alar ligaments, either horizontally or off-axis.

We used several pulse sequences, including fast-spin echo (for the motion video loop) and gradient echo as T1-and T2-weighted images. No three-dimensional gradient sequences with secondary reconstruction were used. In order to characterize the instability patterns of the craniocervical joint, especially in the dens and the alar ligaments, we established two criteria to indicate the local motion of the dens: the left-to-right tumbling of the dens ("dancing dens") and the anterior-to-posterior movement during MRI video documentation (figures 1 and 2). In order to identify instability of the craniocervical junction, it is necessary to keep attention focused on the position of the dens, the surrounding dens capsula, and the dimension of the subarachnoid space during the entire rotational maneuver. The types of ligamentous lesions and the identification of instability were documented with respect to the patient's neurotologic and orthopedic presentation, to the operative treatment, and to the patient's outcome.

We found that 72 patients had documented signs of instability, including ligamentous injuries, fibrous capsule ruptures of C1-C2, or trauma to the dens-related capsula. These patients were referred to a neurosurgeon. Surgery was chosen for 42 patients (10.0%) with the intention of obtaining dorsal occipitocervical stabilization. The duration of time between the MRI evaluation and surgery ranged from 1 week to 1.5 years (mean: 3.5 mo). Under the surgical plan, it was initially decided to insert only the bone transplant between C0 and C2 and fix it with a titanium wire. Rotating movements between C1 and C2 were still possible.

Because pseudarthrosis can occur as a result of a recurrence of instability and clinical symptoms, we decided to insert an additional screw, thereby reintroducing an old method described by Magerl. [16] Essentially, this involved a posterior surgical approach to the craniocervical junction. With fluoroscopic guidance, we placed titanium screws into predrilled holes, starting at the arch of C2 and proceeding in the direction of the lateral mass of C1 (figure 3). We paid careful attention to avoiding the intravertebral artery. We achieved further stabilization from C0 to C2 over the spinous process with cortical and spongy autogenous bone graft, which was harvested from the posterior iliac crest and fixed with titanium wiring though two occipital burr holes to the right and left of midline. Further attachment of the bone graft was achieved with "figure 8" wiring to the spinous process of C1 and C2. With this new technique, we were able to achieve immediate and reliable stability.

During followup, we noticed that there was some instability between C0 and C2 in all patients because of the loss of some of the bone graft. We believed that the primary cause of this was the normal movement of the cervical spine. It was our recommendation that a plate be installed between C0 and C2 with fixation screws inserted into the lateral mass of C2 and into the large surface areas of the occipital bone. Even though this procedure prevents rotational movement of the head either to the left or right, it does guarantee that there will be no future separation of C0-C2, such as was the case with the titanium wire. Nowadays, the operative procedure consists of a complete titanium-plate fixation from the occiput to C2 in combination with the Magerl procedure.

Followup data were obtained from hospital medical records, rehabilitation clinics, outpatient charts, and telephone interviews. The collection of followup data is still in progress.

Results

Among the 420 patients we investigated, 20 (4.8%) had a complete rupture of the alar ligament and 52 (12.4%) had an incomplete rupture (table). Twenty-eight of these patients had a coexisting elongation of the transverse atlantal ligament and a malfunction in the contralateral alar ligament. In addition, some of these patients had rupture signs of either the C1-C2 or C2-C3 fibrous capsule. Such a rupture was visible only on maximum contralateral tilting, which caused a subluxation of C2.

Eighty-one patients (19.3%) had a variation in their intraligamentous signal pattern that was probably caused by granulation tissue, and their MRI videos showed evidence of instability. A total of 102 patients (24.3%) had a signal variation in their alar ligaments, but no demonstrable video instability. Another 158 patients (37.6%) showed no evidence of instability in the craniocervical junction and no loss of signal in the alar ligaments. Four patients (1.0%) had dens variations, and three others (0.7%) had elasticity syndromes, such as Ehlers-Danlos syndrome and Marfan's syndrome. Only 12 of the 52 patients (23.1%) who had an incomplete rupture exhibited a periosteal pathology at the insertion of the alar ligaments on the dens. All 20 patients who had a ruptured alar ligament had intraligamentous signal changes in the transverse ligament.

In the 42 patients who underwent stabilization surgery, almost all symptoms had disappeared by postoperative day 5, and considerable improvement was seen in their equilibrium, especially while riding a bicycle. One patient required a repositioning of a screw prior to discharge because of severe unilateral headache.

In most cases, radiographic followup was performed 6 to 8 weeks following the operation. At the 1-year followup, 34 of the 42 surgical patients (80.9%) maintained successful fusion and alleviation of their sensation of instability. Twenty-five of these patients (59.5%)--all of whom were unemployed before surgery--were able to resume a professional activity. In the eight patients (19.0%) who still had a loss of stability during the second and 14th weeks, we noticed that there were some negative effects of rehabilitation. Six of these patients developed a pseudarthrosis or an osteolysis of their bone graft during the first 3 months after fusion, and three required a repeat operation.

Discussion

This is the first followup study based on fMRI video diagnosis to evaluate patients with craniocervical instability caused by a loss of normal function of the alar ligaments after a nonpenetrating cervical spine trauma. When one alar ligament is injured, the main mechanism of motion restriction--axial rotation--is no longer limited. During trauma, especially rear-end automobile accidents that involve a whiplash mechanism and a rotation component, the ligaments are most vulnerable when the head is initially flexed and rotated. [17]

We do not perform rotational CT as described by Dvorak et al [4] and Penning [18] because patients have only a limited ability to move inside the gantry. Also, the anatomic variants of the alar, dental, and transverse ligaments are not revealed by functional CT. In addition, some patients are unable to bend sideways or perform a physiologic maximum rotation. In patients who have vertebral artery territory insufficiency, MRI evaluation should first exclude intravertebral artery pathology because of the risk of restricted blood flow through the contralateral intravertebral artery during passive rotation to the opposite side. In this situation, we performed standby monitoring to minimize the risk of a drop attack. Through direct controlled posturing of the patient, we were able to recognize patient discomfort early.

A previous study showed that seven of 95 patients (7.4%) who underwent fMRI of the upper cervical spine experienced unilateral changes in vertebral artery blood flow, and some had hypoplasia. [15] We concur with Willauschus et al that there is only a low incidence of complete rupture of ligaments in accident victims (4.8% of our patients). [19] We are also in agreement with Dvorak et al [17] and Dickman et al, [20] who suggested that the mechanism of injury in the upper cervical spine occurs first with a rupture of the alar ligaments; only after a complete separation of the alar ligaments has occurred might there be a rupture of the transverse atlantodental ligament. The alar ligaments might be more vulnerable to separation because they are composed mainly of collagenous fibers and they contain few elastic fibers. [21] They are relatively weak compared with other ligaments. [6] Restoration of the strength of these nonfunctional ligaments has not been reported.

One unresolved diagnostic imaging problem is the alteration of signals in patients who have functionally normal craniocervical ligaments. Based on the observation of incomplete rupture of other ligaments--such as those in the shoulder, knee, and ankle joint [22]--we can anticipate that if scarring forms in the disrupted, over-stretched alar ligament, it would manifest as an inhomogeneous, asymmetrical band mass. However, this has not been established with certainty. Therefore, we should endeavor to improve our understanding of the osseous, ligamentous, capsular, and facet-joint variations at the craniocervical junction. [5,9] Before performing an fMRI study in an open system with controlled movements, physicians must be cognizant of the embryologic vascular anatomy and its variants, as well as the defined elasticity syndromes. In our series of 420 patients, we identified 267 (63.6%) with intact stability; 158 of them (37.6% of the total) had signal continuity of the alar ligaments. Moreover, these 267 patien ts might have experienced a nociceptive input failure between the fibrous capsule, the facet-joint synovial membranes, and the interaction of the muscles of the cervicocephalic region. [2,9]

Before considering invasive stabilization surgery, it is imperative that the surgeon scrutinize the MRI signal-intensity patterns of the alar ligaments in order to clearly identify functional as well as anatomic instability. In order to define instability of the craniocervical junction, attention should be given to the position of the densrelated region and the dimensions of its subarachnoid space during the entire rotational maneuver. [23] However, what is truly at the basis of instability will have to be clarified in prospective, controlled, and coordinated outcomes studies by the investigating physicians and the surgeons. Controlled, double-blind outcomes research would be difficult to design for patients who undergo such invasive surgery.

We must not overlook the fact that a cervical disk herniation might be the cause of severe pain, and that significant alleviation of symptoms can be obtained for most of these patients with anterior cervical fusion. Before operating on the upper cervical spine, the surgeon should obtain a routine MRI and, if possible, an anterior-to-posterior MRI under flexion and extension. [24]

In conclusion, fMRI video analysis is a noninvasive investigation to establish instability of the craniocervical junction, especially of the alar ligaments. Routine evaluation of the extracranial vertebral circulation by MRI angiography is an additional preinvestigative recommendation. The mechanism of alar-craniocervical junction instability might be consistent with a severe rotatory-type trauma to the upper cervical spine, such as might occur in an acceleration-deceleration strain to the neck or head while it is in a rotated and flexed position. This type of injury is different from the more typical whiplash event. Surgery is indicated for established unstable injuries when the patient experiences intractable neck pain that has failed to respond to conservative management. In our study, 42 patients with unstable alar ligament injuries and contralateral joint rupture were stabilized by dorsal fusion of CO-Cl-C2. Functional MRI video was able to identify those patients who experienced instability following a severe, traumatic, soft tissue rupture, and it helped direct the most appropriate neurosurgical intervention. Dorsal fusion can correct the underlying instability of the alaratlantoaxial joint and bring about considerable clinical improvement.

Acknowledgment

We thank Daniela Koch for her excellent technical help and Holger Koch for his assistance in preparing the figures.

From the Functional MRI Unit, Institute of Pediatric and Neuroradiology, Kempten, Germany (Dr. Volle), and the Spine Surgery Outpatient Clinic, Ulm, Germany (Dr. Montazem).

Reprint requests: Eckhard Volle, MD, Functional MRI Unit, Institute of Pediatric and Neuroradiology, Poststrasse 16. Hofapotheke, D-87439, Kempten, Allgau, Germany. Phone: [phone]; fax: [phone]; e-mail: [e-mail]

References

(1.) Dvorak J. Funktionelle Anatomic der oberen Halswirbel-saule miter besonderer Berucksichtigung des Bandapparates. In: Wolff HD, hrsg. Die Sonderstellung des Kopfgelenkbereichs Grundlagen, Klinik, Begutachtung. Berlin, Heidelberg, New York: Springer, 1988:19-46.

(2.) Szpalski M, Gunzburg R, Soeur M, et al. Pharmacologic interventions in whiplash-associated disorders. In: Gunzburg R, Szpalski M, eds. Whiplash Injuries: Current Concepts in Prevention, Diagnosis, and Treatment of the Cervical Whiplash Syndrome. Philadelphia: Lippincott-Raven, 1998:175-81.

(3.) Dvorak J, Panjabi MM. Functional anatomy of the alar ligaments. Spine 1987;12:183-9.

(4.) Dvorak J, Penning L, Hayek J, et al. Functional diagnostics of the cervical spine using computer tomography. Neuroradiology 1988;30:132-7.

(5.) Wen N, Lavaste F, Santin JJ, Lassau JP. Three-dimensional biomechanical properties of the human cervical spine in vitro. II. Analysis of instability after ligamentous injuries. Eur Spine J 1993;2:12-5.

(6.) White AA, Panjabi MM. Clinical Biomechanics of the Spine. 2nd ed. Philadelphia: Lippincott, 1990:85-92.

(7.) Actinnes JA, Hayek J, Panjabi MM, Grob D. The value of functional computed tomography in the evaluation of soft-tissue injury in the upper cervical spine. Eur Spine J 1994;3:98-101.

(8.) Dvorak J, Froeblich D, Penning L, et al. Functional radiographic diagnosis of the cervical spine: Flexion/extension. Spine 1988;13:748-55.

(9.) Wolff HD. Neurophysiologische Aspekte des Bewegungssystems, 3. Aufl. Berlin, Heidelberg, New York: Springer, 1996:65-9.

(10.) Volle E, Kreisler P, Wolff HD, et al. Funktionelle Darstellung der Ligamenta alaria in der Kernspintomographie. Manuelle Medizin 1996;34:9-13.

(11.) Satemus KS. Die Wirbelsaulenuntersuchung im Rabmen der forensischen Obduktion. Beitr Gerichtl Med 1988;46:489-95.

(12.) Nidecker A, Pernus B, Hayek J. Ettlin T. ["Whiplash" injury of the cervical spine: Value of modern diagnostic imaging]. Schweiz Med Wochenschr 1997;127:1643-51.

(13.) Benoist M. Natural evolution and resolution of the cervical whiplash syndrome. In: Gunzburg R, Szpalski M, eds. Whiplash Injuries: Current Concepts in Prevention, Diagnosis, and Treatment of the Cervical Whiplash Syndrome. Philadelphia: Lippincott-Raven, 1998:117-26.

(14.) Sturzenegger M, DiStefano G, Radanov BP, Schnidrig A. Presenting symptoms and signs after whiplash injury: The influence of accident mechanisms. Neurology 1994;44:688-93.

(15.) Volle E, Montazem A. Strukturdefekte der Ligamenta alaria in der offenen Funktionskernspintomographie. Manuelle Medizin 1997;35: 18 8-93.

(16.) Grob D. Posterior surgery. In: Gunzburg R, Szpalski M, eds. Whiplash Injuries: Current Concepts in Prevention, Diagnosis, and Treatment of the Cervical Whiplash Syndrome. Philadelphia: Lippincott-Raven, 1998:241-6.

(17.) Dvorak J, Schneider E, Saldinger P, Rahn B. Biomechanics of the craniocervical region: The alar and transverse ligaments. J Orthop Res 1988;6:452-61.

(18.) Penning L. Acceleration injury of the cervical spine by hypertranslation of the head: Part 2. Effect of hypertranslation of the head on the cervical spine motion: Discussion of literature data. Eur Spine J 1992;l:13-9.

(19.) Willauschus WG, Kladny B, Beyer WF, et al. Lesions of the alar ligaments. In vivo and in vitro studies with magnetic resonance imaging. Spine 1995;20:2493-8.

(20.) Dickman CA, Greene KA, Sonntag VK. Injuries involving the transverse atlantal ligament: Classification and treatment guidelines based upon experience with 39 injuries. Neurosurgery 1996;38:44-50.

(21.) SaldingerP, Dvorak J, Rahn BA, Perren SM. Histology of the alar and transverse ligaments. Spine 1990;15:257-61.

(22.) Stoller DW. Magnetic Resonance Imaging in Orthopaedics and Sports Medicine. 2nd ed. Philadelphia: Lippincott-Raven, 1997.

(23.) Steel H. Anatomical and mechanical consideration of the atlantoaxial articulation. J Bone Joint Surg [Am] 1969;50:1481-7.

(24.) Shellock FG, Sullenberger P, Mink JH, et al. MRI of the cervical spine during flexion and extension: Development and implementation of a new technique. J Magn Reson Imaging 1994;21:712-3.

                  Analysis of MRI and MRI video findings
                  in craniocervical instability pathology
                  of 420 patients with clinical symptoms
                consistent with possible spinal instability
Findings                               Patients
                                        n (%)
Complete alar ligament rupture [*]     20 (4.8)
Incomplete alar ligament rupture, [*]  52 (12.4)
Alar signal-pattern variation;
  instability                          81 (19.3)
Alar signal-pattern variation; no
  instability                         102 (24.3)
Normal alar signal pattern; no
  instability                         158 (37.6)
Dens variation                          4 (1.0)
Elasticity syndromes                    3 (0.7)
Total                                 420 (100)
(*.)These 72 patients (17.1%) were referred to a neurosurgeon. Of them,
42 (10.0%) underwent posterior spinal fusion.